AddressSanitizer.cpp revision 8375bc923638299410e75e8f668524fc1aa113bd 
1//===-- AddressSanitizer.cpp - memory error detector ------------*- C++ -*-===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This file is a part of AddressSanitizer, an address sanity checker.
11// Details of the algorithm:
12//  http://code.google.com/p/address-sanitizer/wiki/AddressSanitizerAlgorithm
13//
14//===----------------------------------------------------------------------===//
15
16#define DEBUG_TYPE "asan"
17
18#include "llvm/ADT/ArrayRef.h"
19#include "llvm/ADT/OwningPtr.h"
20#include "llvm/ADT/SmallSet.h"
21#include "llvm/ADT/SmallString.h"
22#include "llvm/ADT/SmallVector.h"
23#include "llvm/ADT/StringExtras.h"
24#include "llvm/Function.h"
25#include "llvm/IntrinsicInst.h"
26#include "llvm/LLVMContext.h"
27#include "llvm/Module.h"
28#include "llvm/Support/CommandLine.h"
29#include "llvm/Support/DataTypes.h"
30#include "llvm/Support/Debug.h"
31#include "llvm/Support/IRBuilder.h"
32#include "llvm/Support/MemoryBuffer.h"
33#include "llvm/Support/Regex.h"
34#include "llvm/Support/raw_ostream.h"
35#include "llvm/Support/system_error.h"
36#include "llvm/Target/TargetData.h"
37#include "llvm/Target/TargetMachine.h"
38#include "llvm/Transforms/Instrumentation.h"
39#include "llvm/Transforms/Utils/BasicBlockUtils.h"
40#include "llvm/Transforms/Utils/ModuleUtils.h"
41#include "llvm/Type.h"
42
43#include <string>
44#include <algorithm>
45
46using namespace llvm;
47
48static const uint64_t kDefaultShadowScale = 3;
49static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
50static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
51
52static const size_t kMaxStackMallocSize = 1 << 16;  // 64K
53static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
54static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
55
56static const char *kAsanModuleCtorName = "asan.module_ctor";
57static const char *kAsanModuleDtorName = "asan.module_dtor";
58static const int   kAsanCtorAndCtorPriority = 1;
59static const char *kAsanReportErrorTemplate = "__asan_report_";
60static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
61static const char *kAsanUnregisterGlobalsName = "__asan_unregister_globals";
62static const char *kAsanInitName = "__asan_init";
63static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
64static const char *kAsanMappingScaleName = "__asan_mapping_scale";
65static const char *kAsanStackMallocName = "__asan_stack_malloc";
66static const char *kAsanStackFreeName = "__asan_stack_free";
67
68static const int kAsanStackLeftRedzoneMagic = 0xf1;
69static const int kAsanStackMidRedzoneMagic = 0xf2;
70static const int kAsanStackRightRedzoneMagic = 0xf3;
71static const int kAsanStackPartialRedzoneMagic = 0xf4;
72
73// Command-line flags.
74
75// This flag may need to be replaced with -f[no-]asan-reads.
76static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
77       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
78static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
79       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
80// This flag may need to be replaced with -f[no]asan-stack.
81static cl::opt<bool> ClStack("asan-stack",
82       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
83// This flag may need to be replaced with -f[no]asan-use-after-return.
84static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
85       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
86// This flag may need to be replaced with -f[no]asan-globals.
87static cl::opt<bool> ClGlobals("asan-globals",
88       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
89static cl::opt<bool> ClMemIntrin("asan-memintrin",
90       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
91// This flag may need to be replaced with -fasan-blacklist.
92static cl::opt<std::string>  ClBlackListFile("asan-blacklist",
93       cl::desc("File containing the list of functions to ignore "
94                "during instrumentation"), cl::Hidden);
95
96// These flags allow to change the shadow mapping.
97// The shadow mapping looks like
98//    Shadow = (Mem >> scale) + (1 << offset_log)
99static cl::opt<int> ClMappingScale("asan-mapping-scale",
100       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
101static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
102       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
103
104// Optimization flags. Not user visible, used mostly for testing
105// and benchmarking the tool.
106static cl::opt<bool> ClOpt("asan-opt",
107       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
108static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
109       cl::desc("Instrument the same temp just once"), cl::Hidden,
110       cl::init(true));
111static cl::opt<bool> ClOptGlobals("asan-opt-globals",
112       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
113
114// Debug flags.
115static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
116                            cl::init(0));
117static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
118                                 cl::Hidden, cl::init(0));
119static cl::opt<std::string> ClDebugFunc("asan-debug-func",
120                                        cl::Hidden, cl::desc("Debug func"));
121static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
122                               cl::Hidden, cl::init(-1));
123static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
124                               cl::Hidden, cl::init(-1));
125
126namespace {
127
128// Blacklisted functions are not instrumented.
129// The blacklist file contains one or more lines like this:
130// ---
131// fun:FunctionWildCard
132// ---
133// This is similar to the "ignore" feature of ThreadSanitizer.
134// http://code.google.com/p/data-race-test/wiki/ThreadSanitizerIgnores
135class BlackList {
136 public:
137  BlackList(const std::string &Path);
138  bool isIn(const Function &F);
139 private:
140  Regex *Functions;
141};
142
143/// AddressSanitizer: instrument the code in module to find memory bugs.
144struct AddressSanitizer : public ModulePass {
145  AddressSanitizer();
146  virtual const char *getPassName() const;
147  void instrumentMop(Instruction *I);
148  void instrumentAddress(Instruction *OrigIns, IRBuilder<> &IRB,
149                         Value *Addr, uint32_t TypeSize, bool IsWrite);
150  Instruction *generateCrashCode(IRBuilder<> &IRB, Value *Addr,
151                                 bool IsWrite, uint32_t TypeSize);
152  bool instrumentMemIntrinsic(MemIntrinsic *MI);
153  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
154                                  Value *Size,
155                                   Instruction *InsertBefore, bool IsWrite);
156  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
157  bool handleFunction(Module &M, Function &F);
158  bool poisonStackInFunction(Module &M, Function &F);
159  virtual bool runOnModule(Module &M);
160  bool insertGlobalRedzones(Module &M);
161  BranchInst *splitBlockAndInsertIfThen(Instruction *SplitBefore, Value *Cmp);
162  static char ID;  // Pass identification, replacement for typeid
163
164 private:
165
166  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
167    Type *Ty = AI->getAllocatedType();
168    uint64_t SizeInBytes = TD->getTypeStoreSizeInBits(Ty) / 8;
169    return SizeInBytes;
170  }
171  uint64_t getAlignedSize(uint64_t SizeInBytes) {
172    return ((SizeInBytes + RedzoneSize - 1)
173            / RedzoneSize) * RedzoneSize;
174  }
175  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
176    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
177    return getAlignedSize(SizeInBytes);
178  }
179
180  void PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
181                   Value *ShadowBase, bool DoPoison);
182  bool LooksLikeCodeInBug11395(Instruction *I);
183
184  Module      *CurrentModule;
185  LLVMContext *C;
186  TargetData *TD;
187  uint64_t MappingOffset;
188  int MappingScale;
189  size_t RedzoneSize;
190  int LongSize;
191  Type *IntptrTy;
192  Type *IntptrPtrTy;
193  Function *AsanCtorFunction;
194  Function *AsanInitFunction;
195  Instruction *CtorInsertBefore;
196  OwningPtr<BlackList> BL;
197};
198}  // namespace
199
200char AddressSanitizer::ID = 0;
201INITIALIZE_PASS(AddressSanitizer, "asan",
202    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
203    false, false)
204AddressSanitizer::AddressSanitizer() : ModulePass(ID) { }
205ModulePass *llvm::createAddressSanitizerPass() {
206  return new AddressSanitizer();
207}
208
209const char *AddressSanitizer::getPassName() const {
210  return "AddressSanitizer";
211}
212
213// Create a constant for Str so that we can pass it to the run-time lib.
214static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
215  Constant *StrConst = ConstantArray::get(M.getContext(), Str);
216  return new GlobalVariable(M, StrConst->getType(), true,
217                            GlobalValue::PrivateLinkage, StrConst, "");
218}
219
220// Split the basic block and insert an if-then code.
221// Before:
222//   Head
223//   SplitBefore
224//   Tail
225// After:
226//   Head
227//   if (Cmp)
228//     NewBasicBlock
229//   SplitBefore
230//   Tail
231//
232// Returns the NewBasicBlock's terminator.
233BranchInst *AddressSanitizer::splitBlockAndInsertIfThen(
234    Instruction *SplitBefore, Value *Cmp) {
235  BasicBlock *Head = SplitBefore->getParent();
236  BasicBlock *Tail = Head->splitBasicBlock(SplitBefore);
237  TerminatorInst *HeadOldTerm = Head->getTerminator();
238  BasicBlock *NewBasicBlock =
239      BasicBlock::Create(*C, "", Head->getParent());
240  BranchInst *HeadNewTerm = BranchInst::Create(/*ifTrue*/NewBasicBlock,
241                                               /*ifFalse*/Tail,
242                                               Cmp);
243  ReplaceInstWithInst(HeadOldTerm, HeadNewTerm);
244
245  BranchInst *CheckTerm = BranchInst::Create(Tail, NewBasicBlock);
246  return CheckTerm;
247}
248
249Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
250  // Shadow >> scale
251  Shadow = IRB.CreateLShr(Shadow, MappingScale);
252  if (MappingOffset == 0)
253    return Shadow;
254  // (Shadow >> scale) | offset
255  return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
256                                               MappingOffset));
257}
258
259void AddressSanitizer::instrumentMemIntrinsicParam(Instruction *OrigIns,
260    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
261  // Check the first byte.
262  {
263    IRBuilder<> IRB(InsertBefore);
264    instrumentAddress(OrigIns, IRB, Addr, 8, IsWrite);
265  }
266  // Check the last byte.
267  {
268    IRBuilder<> IRB(InsertBefore);
269    Value *SizeMinusOne = IRB.CreateSub(
270        Size, ConstantInt::get(Size->getType(), 1));
271    SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
272    Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
273    Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
274    instrumentAddress(OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
275  }
276}
277
278// Instrument memset/memmove/memcpy
279bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
280  Value *Dst = MI->getDest();
281  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
282  Value *Src = MemTran ? MemTran->getSource() : NULL;
283  Value *Length = MI->getLength();
284
285  Constant *ConstLength = dyn_cast<Constant>(Length);
286  Instruction *InsertBefore = MI;
287  if (ConstLength) {
288    if (ConstLength->isNullValue()) return false;
289  } else {
290    // The size is not a constant so it could be zero -- check at run-time.
291    IRBuilder<> IRB(InsertBefore);
292
293    Value *Cmp = IRB.CreateICmpNE(Length,
294                                   Constant::getNullValue(Length->getType()));
295    InsertBefore = splitBlockAndInsertIfThen(InsertBefore, Cmp);
296  }
297
298  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
299  if (Src)
300    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
301  return true;
302}
303
304static Value *getLDSTOperand(Instruction *I) {
305  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
306    return LI->getPointerOperand();
307  }
308  return cast<StoreInst>(*I).getPointerOperand();
309}
310
311void AddressSanitizer::instrumentMop(Instruction *I) {
312  int IsWrite = isa<StoreInst>(*I);
313  Value *Addr = getLDSTOperand(I);
314  if (ClOpt && ClOptGlobals && isa<GlobalVariable>(Addr)) {
315    // We are accessing a global scalar variable. Nothing to catch here.
316    return;
317  }
318  Type *OrigPtrTy = Addr->getType();
319  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
320
321  assert(OrigTy->isSized());
322  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
323
324  if (TypeSize != 8  && TypeSize != 16 &&
325      TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
326    // Ignore all unusual sizes.
327    return;
328  }
329
330  IRBuilder<> IRB(I);
331  instrumentAddress(I, IRB, Addr, TypeSize, IsWrite);
332}
333
334Instruction *AddressSanitizer::generateCrashCode(
335    IRBuilder<> &IRB, Value *Addr, bool IsWrite, uint32_t TypeSize) {
336  // IsWrite and TypeSize are encoded in the function name.
337  std::string FunctionName = std::string(kAsanReportErrorTemplate) +
338      (IsWrite ? "store" : "load") + itostr(TypeSize / 8);
339  Value *ReportWarningFunc = CurrentModule->getOrInsertFunction(
340      FunctionName, IRB.getVoidTy(), IntptrTy, NULL);
341  CallInst *Call = IRB.CreateCall(ReportWarningFunc, Addr);
342  Call->setDoesNotReturn();
343  return Call;
344}
345
346void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
347                                         IRBuilder<> &IRB, Value *Addr,
348                                         uint32_t TypeSize, bool IsWrite) {
349  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
350
351  Type *ShadowTy  = IntegerType::get(
352      *C, std::max(8U, TypeSize >> MappingScale));
353  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
354  Value *ShadowPtr = memToShadow(AddrLong, IRB);
355  Value *CmpVal = Constant::getNullValue(ShadowTy);
356  Value *ShadowValue = IRB.CreateLoad(
357      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
358
359  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
360
361  Instruction *CheckTerm = splitBlockAndInsertIfThen(
362      cast<Instruction>(Cmp)->getNextNode(), Cmp);
363  IRBuilder<> IRB2(CheckTerm);
364
365  size_t Granularity = 1 << MappingScale;
366  if (TypeSize < 8 * Granularity) {
367    // Addr & (Granularity - 1)
368    Value *Lower3Bits = IRB2.CreateAnd(
369        AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
370    // (Addr & (Granularity - 1)) + size - 1
371    Value *LastAccessedByte = IRB2.CreateAdd(
372        Lower3Bits, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
373    // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
374    LastAccessedByte = IRB2.CreateIntCast(
375        LastAccessedByte, IRB.getInt8Ty(), false);
376    // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
377    Value *Cmp2 = IRB2.CreateICmpSGE(LastAccessedByte, ShadowValue);
378
379    CheckTerm = splitBlockAndInsertIfThen(CheckTerm, Cmp2);
380  }
381
382  IRBuilder<> IRB1(CheckTerm);
383  Instruction *Crash = generateCrashCode(IRB1, AddrLong, IsWrite, TypeSize);
384  Crash->setDebugLoc(OrigIns->getDebugLoc());
385  ReplaceInstWithInst(CheckTerm, new UnreachableInst(*C));
386}
387
388// This function replaces all global variables with new variables that have
389// trailing redzones. It also creates a function that poisons
390// redzones and inserts this function into llvm.global_ctors.
391bool AddressSanitizer::insertGlobalRedzones(Module &M) {
392  SmallVector<GlobalVariable *, 16> GlobalsToChange;
393
394  for (Module::GlobalListType::iterator G = M.getGlobalList().begin(),
395       E = M.getGlobalList().end(); G != E; ++G) {
396    Type *Ty = cast<PointerType>(G->getType())->getElementType();
397    DEBUG(dbgs() << "GLOBAL: " << *G);
398
399    if (!Ty->isSized()) continue;
400    if (!G->hasInitializer()) continue;
401    // Touch only those globals that will not be defined in other modules.
402    // Don't handle ODR type linkages since other modules may be built w/o asan.
403    if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
404        G->getLinkage() != GlobalVariable::PrivateLinkage &&
405        G->getLinkage() != GlobalVariable::InternalLinkage)
406      continue;
407    // Two problems with thread-locals:
408    //   - The address of the main thread's copy can't be computed at link-time.
409    //   - Need to poison all copies, not just the main thread's one.
410    if (G->isThreadLocal())
411      continue;
412    // For now, just ignore this Alloca if the alignment is large.
413    if (G->getAlignment() > RedzoneSize) continue;
414
415    // Ignore all the globals with the names starting with "\01L_OBJC_".
416    // Many of those are put into the .cstring section. The linker compresses
417    // that section by removing the spare \0s after the string terminator, so
418    // our redzones get broken.
419    if ((G->getName().find("\01L_OBJC_") == 0) ||
420        (G->getName().find("\01l_OBJC_") == 0)) {
421      DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
422      continue;
423    }
424
425    if (G->hasSection()) {
426      StringRef Section(G->getSection());
427      // Ignore the globals from the __OBJC section. The ObjC runtime assumes
428      // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
429      // them.
430      if ((Section.find("__OBJC,") == 0) ||
431          (Section.find("__DATA, __objc_") == 0)) {
432        DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
433        continue;
434      }
435      // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
436      // Constant CFString instances are compiled in the following way:
437      //  -- the string buffer is emitted into
438      //     __TEXT,__cstring,cstring_literals
439      //  -- the constant NSConstantString structure referencing that buffer
440      //     is placed into __DATA,__cfstring
441      // Therefore there's no point in placing redzones into __DATA,__cfstring.
442      // Moreover, it causes the linker to crash on OS X 10.7
443      if (Section.find("__DATA,__cfstring") == 0) {
444        DEBUG(dbgs() << "Ignoring CFString: " << *G);
445        continue;
446      }
447    }
448
449    GlobalsToChange.push_back(G);
450  }
451
452  size_t n = GlobalsToChange.size();
453  if (n == 0) return false;
454
455  // A global is described by a structure
456  //   size_t beg;
457  //   size_t size;
458  //   size_t size_with_redzone;
459  //   const char *name;
460  // We initialize an array of such structures and pass it to a run-time call.
461  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
462                                               IntptrTy, IntptrTy, NULL);
463  SmallVector<Constant *, 16> Initializers(n);
464
465  IRBuilder<> IRB(CtorInsertBefore);
466
467  for (size_t i = 0; i < n; i++) {
468    GlobalVariable *G = GlobalsToChange[i];
469    PointerType *PtrTy = cast<PointerType>(G->getType());
470    Type *Ty = PtrTy->getElementType();
471    uint64_t SizeInBytes = TD->getTypeStoreSizeInBits(Ty) / 8;
472    uint64_t RightRedzoneSize = RedzoneSize +
473        (RedzoneSize - (SizeInBytes % RedzoneSize));
474    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
475
476    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
477    Constant *NewInitializer = ConstantStruct::get(
478        NewTy, G->getInitializer(),
479        Constant::getNullValue(RightRedZoneTy), NULL);
480
481    SmallString<2048> DescriptionOfGlobal = G->getName();
482    DescriptionOfGlobal += " (";
483    DescriptionOfGlobal += M.getModuleIdentifier();
484    DescriptionOfGlobal += ")";
485    GlobalVariable *Name = createPrivateGlobalForString(M, DescriptionOfGlobal);
486
487    // Create a new global variable with enough space for a redzone.
488    GlobalVariable *NewGlobal = new GlobalVariable(
489        M, NewTy, G->isConstant(), G->getLinkage(),
490        NewInitializer, "", G, G->isThreadLocal());
491    NewGlobal->copyAttributesFrom(G);
492    NewGlobal->setAlignment(RedzoneSize);
493
494    Value *Indices2[2];
495    Indices2[0] = IRB.getInt32(0);
496    Indices2[1] = IRB.getInt32(0);
497
498    G->replaceAllUsesWith(
499        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
500    NewGlobal->takeName(G);
501    G->eraseFromParent();
502
503    Initializers[i] = ConstantStruct::get(
504        GlobalStructTy,
505        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
506        ConstantInt::get(IntptrTy, SizeInBytes),
507        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
508        ConstantExpr::getPointerCast(Name, IntptrTy),
509        NULL);
510    DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
511  }
512
513  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
514  GlobalVariable *AllGlobals = new GlobalVariable(
515      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
516      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
517
518  Function *AsanRegisterGlobals = cast<Function>(M.getOrInsertFunction(
519      kAsanRegisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
520  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
521
522  IRB.CreateCall2(AsanRegisterGlobals,
523                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
524                  ConstantInt::get(IntptrTy, n));
525
526  // We also need to unregister globals at the end, e.g. when a shared library
527  // gets closed.
528  Function *AsanDtorFunction = Function::Create(
529      FunctionType::get(Type::getVoidTy(*C), false),
530      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
531  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
532  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
533  Function *AsanUnregisterGlobals = cast<Function>(M.getOrInsertFunction(
534      kAsanUnregisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
535  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
536
537  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
538                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
539                       ConstantInt::get(IntptrTy, n));
540  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
541
542  DEBUG(dbgs() << M);
543  return true;
544}
545
546// virtual
547bool AddressSanitizer::runOnModule(Module &M) {
548  // Initialize the private fields. No one has accessed them before.
549  TD = getAnalysisIfAvailable<TargetData>();
550  if (!TD)
551    return false;
552  BL.reset(new BlackList(ClBlackListFile));
553
554  CurrentModule = &M;
555  C = &(M.getContext());
556  LongSize = TD->getPointerSizeInBits();
557  IntptrTy = Type::getIntNTy(*C, LongSize);
558  IntptrPtrTy = PointerType::get(IntptrTy, 0);
559
560  AsanCtorFunction = Function::Create(
561      FunctionType::get(Type::getVoidTy(*C), false),
562      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
563  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
564  CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
565
566  // call __asan_init in the module ctor.
567  IRBuilder<> IRB(CtorInsertBefore);
568  AsanInitFunction = cast<Function>(
569      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
570  AsanInitFunction->setLinkage(Function::ExternalLinkage);
571  IRB.CreateCall(AsanInitFunction);
572
573  MappingOffset = LongSize == 32
574      ? kDefaultShadowOffset32 : kDefaultShadowOffset64;
575  if (ClMappingOffsetLog >= 0) {
576    if (ClMappingOffsetLog == 0) {
577      // special case
578      MappingOffset = 0;
579    } else {
580      MappingOffset = 1ULL << ClMappingOffsetLog;
581    }
582  }
583  MappingScale = kDefaultShadowScale;
584  if (ClMappingScale) {
585    MappingScale = ClMappingScale;
586  }
587  // Redzone used for stack and globals is at least 32 bytes.
588  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
589  RedzoneSize = std::max(32, (int)(1 << MappingScale));
590
591  bool Res = false;
592
593  if (ClGlobals)
594    Res |= insertGlobalRedzones(M);
595
596  // Tell the run-time the current values of mapping offset and scale.
597  GlobalValue *asan_mapping_offset =
598      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
599                     ConstantInt::get(IntptrTy, MappingOffset),
600                     kAsanMappingOffsetName);
601  GlobalValue *asan_mapping_scale =
602      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
603                         ConstantInt::get(IntptrTy, MappingScale),
604                         kAsanMappingScaleName);
605  // Read these globals, otherwise they may be optimized away.
606  IRB.CreateLoad(asan_mapping_scale, true);
607  IRB.CreateLoad(asan_mapping_offset, true);
608
609
610  for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
611    if (F->isDeclaration()) continue;
612    Res |= handleFunction(M, *F);
613  }
614
615  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
616
617  return Res;
618}
619
620bool AddressSanitizer::handleFunction(Module &M, Function &F) {
621  if (BL->isIn(F)) return false;
622  if (&F == AsanCtorFunction) return false;
623  if (!F.hasFnAttr(Attribute::AddressSafety)) return false;
624
625  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
626    return false;
627  // We want to instrument every address only once per basic block
628  // (unless there are calls between uses).
629  SmallSet<Value*, 16> TempsToInstrument;
630  SmallVector<Instruction*, 16> ToInstrument;
631
632  // Fill the set of memory operations to instrument.
633  for (Function::iterator FI = F.begin(), FE = F.end();
634       FI != FE; ++FI) {
635    TempsToInstrument.clear();
636    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
637         BI != BE; ++BI) {
638      if (LooksLikeCodeInBug11395(BI)) return false;
639      if ((isa<LoadInst>(BI) && ClInstrumentReads) ||
640          (isa<StoreInst>(BI) && ClInstrumentWrites)) {
641        Value *Addr = getLDSTOperand(BI);
642        if (ClOpt && ClOptSameTemp) {
643          if (!TempsToInstrument.insert(Addr))
644            continue;  // We've seen this temp in the current BB.
645        }
646      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
647        // ok, take it.
648      } else {
649        if (isa<CallInst>(BI)) {
650          // A call inside BB.
651          TempsToInstrument.clear();
652        }
653        continue;
654      }
655      ToInstrument.push_back(BI);
656    }
657  }
658
659  // Instrument.
660  int NumInstrumented = 0;
661  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
662    Instruction *Inst = ToInstrument[i];
663    if (ClDebugMin < 0 || ClDebugMax < 0 ||
664        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
665      if (isa<StoreInst>(Inst) || isa<LoadInst>(Inst))
666        instrumentMop(Inst);
667      else
668        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
669    }
670    NumInstrumented++;
671  }
672
673  DEBUG(dbgs() << F);
674
675  bool ChangedStack = poisonStackInFunction(M, F);
676
677  // For each NSObject descendant having a +load method, this method is invoked
678  // by the ObjC runtime before any of the static constructors is called.
679  // Therefore we need to instrument such methods with a call to __asan_init
680  // at the beginning in order to initialize our runtime before any access to
681  // the shadow memory.
682  // We cannot just ignore these methods, because they may call other
683  // instrumented functions.
684  if (F.getName().find(" load]") != std::string::npos) {
685    IRBuilder<> IRB(F.begin()->begin());
686    IRB.CreateCall(AsanInitFunction);
687  }
688
689  return NumInstrumented > 0 || ChangedStack;
690}
691
692static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
693  if (ShadowRedzoneSize == 1) return PoisonByte;
694  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
695  if (ShadowRedzoneSize == 4)
696    return (PoisonByte << 24) + (PoisonByte << 16) +
697        (PoisonByte << 8) + (PoisonByte);
698  assert(0 && "ShadowRedzoneSize is either 1, 2 or 4");
699  return 0;
700}
701
702static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
703                                            size_t Size,
704                                            size_t RedzoneSize,
705                                            size_t ShadowGranularity,
706                                            uint8_t Magic) {
707  for (size_t i = 0; i < RedzoneSize;
708       i+= ShadowGranularity, Shadow++) {
709    if (i + ShadowGranularity <= Size) {
710      *Shadow = 0;  // fully addressable
711    } else if (i >= Size) {
712      *Shadow = Magic;  // unaddressable
713    } else {
714      *Shadow = Size - i;  // first Size-i bytes are addressable
715    }
716  }
717}
718
719void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
720                                   IRBuilder<> IRB,
721                                   Value *ShadowBase, bool DoPoison) {
722  size_t ShadowRZSize = RedzoneSize >> MappingScale;
723  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
724  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
725  Type *RZPtrTy = PointerType::get(RZTy, 0);
726
727  Value *PoisonLeft  = ConstantInt::get(RZTy,
728    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
729  Value *PoisonMid   = ConstantInt::get(RZTy,
730    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
731  Value *PoisonRight = ConstantInt::get(RZTy,
732    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
733
734  // poison the first red zone.
735  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
736
737  // poison all other red zones.
738  uint64_t Pos = RedzoneSize;
739  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
740    AllocaInst *AI = AllocaVec[i];
741    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
742    uint64_t AlignedSize = getAlignedAllocaSize(AI);
743    assert(AlignedSize - SizeInBytes < RedzoneSize);
744    Value *Ptr = NULL;
745
746    Pos += AlignedSize;
747
748    assert(ShadowBase->getType() == IntptrTy);
749    if (SizeInBytes < AlignedSize) {
750      // Poison the partial redzone at right
751      Ptr = IRB.CreateAdd(
752          ShadowBase, ConstantInt::get(IntptrTy,
753                                       (Pos >> MappingScale) - ShadowRZSize));
754      size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
755      uint32_t Poison = 0;
756      if (DoPoison) {
757        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
758                                        RedzoneSize,
759                                        1ULL << MappingScale,
760                                        kAsanStackPartialRedzoneMagic);
761      }
762      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
763      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
764    }
765
766    // Poison the full redzone at right.
767    Ptr = IRB.CreateAdd(ShadowBase,
768                        ConstantInt::get(IntptrTy, Pos >> MappingScale));
769    Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
770    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
771
772    Pos += RedzoneSize;
773  }
774}
775
776// Workaround for bug 11395: we don't want to instrument stack in functions
777// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
778// FIXME: remove once the bug 11395 is fixed.
779bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
780  if (LongSize != 32) return false;
781  CallInst *CI = dyn_cast<CallInst>(I);
782  if (!CI || !CI->isInlineAsm()) return false;
783  if (CI->getNumArgOperands() <= 5) return false;
784  // We have inline assembly with quite a few arguments.
785  return true;
786}
787
788// Find all static Alloca instructions and put
789// poisoned red zones around all of them.
790// Then unpoison everything back before the function returns.
791//
792// Stack poisoning does not play well with exception handling.
793// When an exception is thrown, we essentially bypass the code
794// that unpoisones the stack. This is why the run-time library has
795// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
796// stack in the interceptor. This however does not work inside the
797// actual function which catches the exception. Most likely because the
798// compiler hoists the load of the shadow value somewhere too high.
799// This causes asan to report a non-existing bug on 453.povray.
800// It sounds like an LLVM bug.
801bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
802  if (!ClStack) return false;
803  SmallVector<AllocaInst*, 16> AllocaVec;
804  SmallVector<Instruction*, 8> RetVec;
805  uint64_t TotalSize = 0;
806
807  // Filter out Alloca instructions we want (and can) handle.
808  // Collect Ret instructions.
809  for (Function::iterator FI = F.begin(), FE = F.end();
810       FI != FE; ++FI) {
811    BasicBlock &BB = *FI;
812    for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
813         BI != BE; ++BI) {
814      if (isa<ReturnInst>(BI)) {
815          RetVec.push_back(BI);
816          continue;
817      }
818
819      AllocaInst *AI = dyn_cast<AllocaInst>(BI);
820      if (!AI) continue;
821      if (AI->isArrayAllocation()) continue;
822      if (!AI->isStaticAlloca()) continue;
823      if (!AI->getAllocatedType()->isSized()) continue;
824      if (AI->getAlignment() > RedzoneSize) continue;
825      AllocaVec.push_back(AI);
826      uint64_t AlignedSize =  getAlignedAllocaSize(AI);
827      TotalSize += AlignedSize;
828    }
829  }
830
831  if (AllocaVec.empty()) return false;
832
833  uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
834
835  bool DoStackMalloc = ClUseAfterReturn
836      && LocalStackSize <= kMaxStackMallocSize;
837
838  Instruction *InsBefore = AllocaVec[0];
839  IRBuilder<> IRB(InsBefore);
840
841
842  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
843  AllocaInst *MyAlloca =
844      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
845  MyAlloca->setAlignment(RedzoneSize);
846  assert(MyAlloca->isStaticAlloca());
847  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
848  Value *LocalStackBase = OrigStackBase;
849
850  if (DoStackMalloc) {
851    Value *AsanStackMallocFunc = M.getOrInsertFunction(
852        kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
853    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
854        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
855  }
856
857  // This string will be parsed by the run-time (DescribeStackAddress).
858  SmallString<2048> StackDescriptionStorage;
859  raw_svector_ostream StackDescription(StackDescriptionStorage);
860  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
861
862  uint64_t Pos = RedzoneSize;
863  // Replace Alloca instructions with base+offset.
864  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
865    AllocaInst *AI = AllocaVec[i];
866    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
867    StringRef Name = AI->getName();
868    StackDescription << Pos << " " << SizeInBytes << " "
869                     << Name.size() << " " << Name << " ";
870    uint64_t AlignedSize = getAlignedAllocaSize(AI);
871    assert((AlignedSize % RedzoneSize) == 0);
872    AI->replaceAllUsesWith(
873        IRB.CreateIntToPtr(
874            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
875            AI->getType()));
876    Pos += AlignedSize + RedzoneSize;
877  }
878  assert(Pos == LocalStackSize);
879
880  // Write the Magic value and the frame description constant to the redzone.
881  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
882  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
883                  BasePlus0);
884  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
885                                   ConstantInt::get(IntptrTy, LongSize/8));
886  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
887  Value *Description = IRB.CreatePointerCast(
888      createPrivateGlobalForString(M, StackDescription.str()),
889      IntptrTy);
890  IRB.CreateStore(Description, BasePlus1);
891
892  // Poison the stack redzones at the entry.
893  Value *ShadowBase = memToShadow(LocalStackBase, IRB);
894  PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
895
896  Value *AsanStackFreeFunc = NULL;
897  if (DoStackMalloc) {
898    AsanStackFreeFunc = M.getOrInsertFunction(
899        kAsanStackFreeName, IRB.getVoidTy(),
900        IntptrTy, IntptrTy, IntptrTy, NULL);
901  }
902
903  // Unpoison the stack before all ret instructions.
904  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
905    Instruction *Ret = RetVec[i];
906    IRBuilder<> IRBRet(Ret);
907
908    // Mark the current frame as retired.
909    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
910                       BasePlus0);
911    // Unpoison the stack.
912    PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
913
914    if (DoStackMalloc) {
915      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
916                         ConstantInt::get(IntptrTy, LocalStackSize),
917                         OrigStackBase);
918    }
919  }
920
921  if (ClDebugStack) {
922    DEBUG(dbgs() << F);
923  }
924
925  return true;
926}
927
928BlackList::BlackList(const std::string &Path) {
929  Functions = NULL;
930  const char *kFunPrefix = "fun:";
931  if (!ClBlackListFile.size()) return;
932  std::string Fun;
933
934  OwningPtr<MemoryBuffer> File;
935  if (error_code EC = MemoryBuffer::getFile(ClBlackListFile.c_str(), File)) {
936    report_fatal_error("Can't open blacklist file " + ClBlackListFile + ": " +
937                       EC.message());
938  }
939  MemoryBuffer *Buff = File.take();
940  const char *Data = Buff->getBufferStart();
941  size_t DataLen = Buff->getBufferSize();
942  SmallVector<StringRef, 16> Lines;
943  SplitString(StringRef(Data, DataLen), Lines, "\n\r");
944  for (size_t i = 0, numLines = Lines.size(); i < numLines; i++) {
945    if (Lines[i].startswith(kFunPrefix)) {
946      std::string ThisFunc = Lines[i].substr(strlen(kFunPrefix));
947      std::string ThisFuncRE;
948      // add ThisFunc replacing * with .*
949      for (size_t j = 0, n = ThisFunc.size(); j < n; j++) {
950        if (ThisFunc[j] == '*')
951          ThisFuncRE += '.';
952        ThisFuncRE += ThisFunc[j];
953      }
954      // Check that the regexp is valid.
955      Regex CheckRE(ThisFuncRE);
956      std::string Error;
957      if (!CheckRE.isValid(Error))
958        report_fatal_error("malformed blacklist regex: " + ThisFunc +
959                           ": " + Error);
960      // Append to the final regexp.
961      if (Fun.size())
962        Fun += "|";
963      Fun += ThisFuncRE;
964    }
965  }
966  if (Fun.size()) {
967    Functions = new Regex(Fun);
968  }
969}
970
971bool BlackList::isIn(const Function &F) {
972  if (Functions) {
973    bool Res = Functions->match(F.getName());
974    return Res;
975  }
976  return false;
977}
978